Solving the Problem of the Accents for Speech Recognition Systems
Irakli Kardava1,
Jemal Antidze1,2, and
Nana Gulua1,2
1. Sokhumi State University, Politkovskaia 9, Tbilisi, 0186, Georgia
2. I. Vekua Scientific Institute of Applied Mathematics, I. Javakhishvili Tbilisi State University, Tbilisi, 0186, Georgia
2. I. Vekua Scientific Institute of Applied Mathematics, I. Javakhishvili Tbilisi State University, Tbilisi, 0186, Georgia
Abstract—Since the speech recognition system has been created, it has developed significantly, but it still has a lot of problems. As you know, any specific natural language may owns about tens accents. Despite the identical word phonemic composition, if it is pronounced in different accents, as a result, we will have sound waves, which are different from each other. Differences in pronunciation, in accent and intonation of speech in general, create one of the most common problems of speech recognition. If there are a lot of accents in language we should create the acoustic model for each separately. When the word is pronounced differently, then the software can become confused and misunderstand (perception) also correctly what is pronounced. The same can also occur, if the human speaks slowly or vice versa quickly, then the program expects. There are any partial decisions (solutions) but they don’t solve all problems. We have developed an approach, which is used to solve above mentioned problems and create more effective, improved speech recognition system of Georgian language and of languages, which are similar to Georgian language. In addition, by the realization of this method, it is available to solve the artificial intelligence issues, such as arrange sound dialogue between computer and human, independent from any accents of any languages.
Index Terms—accents, acoustic model, phonemic composition, speech recognizer, waves
Cite: Irakli Kardava, Jemal Antidze, and Nana Gulua, "Solving the Problem of the Accents for Speech Recognition Systems," International Journal of Signal Processing Systems, Vol. 4, No. 3, pp. 235-238, June 2016. doi: 10.18178/ijsps.4.3.235-238
Cite: Irakli Kardava, Jemal Antidze, and Nana Gulua, "Solving the Problem of the Accents for Speech Recognition Systems," International Journal of Signal Processing Systems, Vol. 4, No. 3, pp. 235-238, June 2016. doi: 10.18178/ijsps.4.3.235-238